Regression
An ML task that predicts a continuous number — sales, salary, temperature.
What is Regression?
Where classification predicts a category, regression predicts a number. Predicting next month's sales (₹2.4 Crore), next-quarter inventory needs (847 units), or a customer's lifetime value (₹14,200) are all regression problems.
Common algorithms: linear regression, gradient boosting (XGBoost, LightGBM), neural networks for complex patterns. Time-series regression is its own subfield with specialised methods (ARIMA, Prophet, transformer-based forecasters).
Evaluation uses different metrics than classification: **MAE** (mean absolute error), **RMSE** (root mean squared error), **MAPE** (mean absolute percentage error). Each penalises different kinds of mistakes — choose based on what matters to the business.
Forecasting + value prediction underpin demand planning, pricing, lifetime-value calculation. Indian retail + FMCG + finance teams hire heavily for this skill.
A Bangalore quick-commerce app predicts demand for 8,000 SKUs in 200 dark stores 6 hours ahead, using regression on weather, time-of-day, and historical patterns. Better forecasting cut spoilage 22%.
Want to master this?
Learn Regression in a structured cohort
3-month live program with mentors, real projects, and 50+ partner placement support.